Estimating sexual size dimorphism in fossil species from posterior probability densities

The evolutionary study of sexual dimorphism has been hampered by the general inability to accurately estimate low levels of dimorphism in fossil assemblages with no information on sex membership. In previous estimates, accuracy decreased with lower dimorphism levels because the hidden distributions of the two sexes embedded in the sample were not adequately modeled. The posterior density peak (pdPeak) method that we developed in part solves this problem by estimating dimorphism and within-sex variance simultaneously and provides unbiased estimates of dimorphism at levels lower than previously possible. Thus, this method has potential to better address the evolutionary history of weakly dimorphic structures, such as the human canine, enabling a better understanding of dimorphisms and adaptive strategies among primate taxa.